Master Data Management
Create Golden Records Every System Can Trust
Profile, match, steward and publish master data so services, finance and operations work from the same facts.
Master Data Management combines data quality measurement, matching and survivorship with stewardship workflows and distribution through APIs, CDC streams and a data marketplace.
- Source records
- Profile
- Match
- Survivorship
- Golden record
- Publish
The enterprise challenge
The same customer, supplier or asset exists five times
Each system holds its own version. Reports disagree, integrations propagate errors and nobody owns the correction.
Master data fails quietly: a duplicated supplier here, a conflicting cost centre there, until reports disagree and automations act on the wrong facts. Master Data Management makes trust explicit. Quality is measured before records are merged, matching proposes candidates with a confidence score, survivorship rules decide which value wins for each attribute, stewards decide the cases rules cannot, and every value in the golden record carries its lineage — so any consumer can ask where a fact came from and get an answer.
Conflicting reports
Finance and operations count suppliers differently.
Silent duplication
New records are created instead of matched.
No lineage
Nobody can say where a value came from.
Signature view
The Golden Record Confluence
Duplicate and conflicting records are resolved into governed golden records with visible lineage.
The Golden Record Confluence: records from several source systems are profiled and matched into one golden record, each attribute tracing back to the source and rule it survived from, with low-confidence matches routed to a steward. Select an attribute to follow its lineage. Illustrative records only, not a product screen.
Tax identifier — Survives from the finance system of record.
- Survivorship
- Rule: system-of-record priority
- Who decides
- Survivorship rule
- Lineage
- Lineage: ERP / finance record and version
What Master Data Management covers
Approved baseline capabilities and the detail restored from the historical module record, labelled by provenance.
Approved baseline
- Approved baseline
Data quality
Measure before you fix.
- Data quality scorecards
- Data profiling
- Quality rules
- Approved baseline
Matching & golden records
Resolve entities.
- AI matching
- Deduplication
- Survivorship
- Golden records
- Hierarchies
- Approved baseline
Stewardship & lineage
Govern the data.
- Data stewardship workflows
- Data lineage
- Approved baseline
Integration & exchange
Distribute trusted data.
- ERP connectivity
- APIs and data exchange
- CDC streams
- Data marketplace capabilities
Restored from the historical record
- Restored · historical record
Data Governance Framework
Establish enterprise data governance for IT operations.
- Data ownership and stewardship assignment
- Data classification and sensitivity labelling
- Governance policy definition and enforcement
- Regulatory obligation mapping (for example, GDPR)
- Data lifecycle management rules
- Restored · historical record
Data Quality Management
Clean, reliable data across all systems.
- Automated data quality scoring
- Duplicate detection and merging
- Validation rules and constraints
- Data cleansing workflows
- Quality trend monitoring and reporting
- Restored · historical record
Data Integration & Synchronization
Unify data across the IT tool ecosystem.
- Bi-directional sync with external systems
- ETL/ELT pipeline management
- Conflict resolution rules
- Streaming and batch synchronization
- API-based data exchange
- Restored · historical record
Reference Data Management
Standardize reference data across the organization.
- Centralized reference data repository
- Category and classification taxonomies
- Location, department and organization hierarchies
- Versioned reference data with change history
- Cross-system reference data alignment
Detail reconciled from the approved historical Nexus module record (additive restoration; legacy performance figures excluded per the claims policy).
From duplicate records to a published golden record
Trigger: New or changed records arrive from a source system.
Ingest
Records received.
Profile
Quality measured.
Match
Candidates scored.
Steward
Exceptions reviewed.
Publish
Golden record distributed.
- Participants
- Data steward
- Data owner
- Consuming system owners
- Work items
- Profiled records
- Match candidates
- Golden record
- Stewardship task
- Automation
- Profiling and rule checks
- Match scoring
- Change publication via CDC
- Governance controls
- Steward approval for low-confidence matches
- Lineage per attribute
Outcome: Consumers receive one governed record with traceable origin.
Match and merge decisions
Match candidates are scored; configured rules handle confident matches and route uncertain ones to a steward, who merges, keeps separate or requests source correction — each outcome recorded.
Survivorship by attribute
Survivorship is defined per attribute — system-of-record priority, most recent validated value or most trusted source — so a golden record can take its tax identifier from finance and its address from the most recently validated source.
Stewardship workload
Stewardship tasks are queued by domain and owner, with both candidate records shown side by side and the rule that flagged them, so stewards decide with context instead of searching.
Distribution with change control
Golden records are published through APIs, CDC streams and data-marketplace subscriptions; changes are logged, so consumers can see what changed and why.
Quality as a measured trend
Scorecards track quality by domain over time, turning data quality from an occasional clean-up into a measured, owned trend reported through Analytics Hub.
Who decides, and what happens when the path breaks
Every decision in the lifecycle has an accountable owner; automation and AI contribute the support listed, never the decision itself.
Decision rights
Merge or keep separate
- Decides
- Data steward
- Support
- Match candidates scored and proposed (recommendation)
Survivorship of conflicting values
- Decides
- Data owner
- Support
- Configured survivorship rules applied (workflow-triggered)
New quality rules
- Decides
- Data owner
- Support
- Rule candidates suggested from profiling (recommendation)
Data classification
- Decides
- Data owner
- Support
- Suggested classification and tagging (recommendation)
Exception register
Match confidence below the configured threshold
- Handling
- Stewardship task created with both records, the match evidence and the rule that flagged it.
- Owner
- Data steward
Conflicting values with no survivorship rule
- Handling
- Routed to the data owner for a decision, which can become a new rule.
- Owner
- Data owner
A merge is found to be wrong
- Handling
- Unmerge using the logged lineage; consumers are notified of the corrected record.
- Owner
- Data steward
A source repeatedly fails quality rules
- Handling
- Quality issue raised to the source system owner with the failing rule and samples.
- Owner
- Source system owner
Stewardship
Named stewards and owners per domain.
Lineage
Origin of each value is recorded.
Change control
Golden-record changes are logged.
Tenant isolation
Records are scoped to the organization's own environment.
Master Data Management in practice
Representative situations drawn from common enterprise security, governance and data patterns.
The supplier who exists five times
The situation
Procurement, finance and the contract repository each hold their own version of the same supplier, with different addresses and payment terms.
With Nexus
Matching proposes the five records as candidates for one entity; survivorship rules take each attribute from its most trusted source, and the resulting golden supplier record is shared with Vendor Management and finance.
The report that cannot be reconciled
The situation
Two dashboards count active cost centres differently because each reads a different system's reference data.
With Nexus
Reference data is managed centrally with versions and change history; both consumers subscribe to the same governed hierarchy, so reports agree because their inputs agree.
The merge nobody can explain
The situation
An automated merge combines two genuinely different organizations, and nobody can tell which rule did it or when.
With Nexus
Low-confidence matches are routed to a steward instead of merged automatically, and every golden-record change is logged with its rule or steward decision — so a wrong merge is traceable and reversible.
Mapped to recognized practice, capability by capability
Each mapping names the Nexus capability and the record that supports the practice. Mappings describe alignment of practice; they are not attestations.
DAMA-DMBOK2
Aligned practice- Practice
- Data Governance, Data Quality and Reference & Master Data knowledge areas
- Nexus capability
- Stewardship workflows, named owners per domain, quality rules, golden-record publishing and reference data management
- Supporting record
- Golden record, stewardship task and data-domain ownership records
ISO 8000 series
Aligned practice- Practice
- Data quality and master data quality characteristics
- Nexus capability
- Profiling, validation rules, duplicate detection and survivorship rules
- Supporting record
- Profiling results and per-attribute survivorship lineage
ISO/IEC 25012:2008
Aligned practice- Practice
- Data quality model — completeness, consistency, accuracy and credibility characteristics
- Nexus capability
- Data quality scorecards and quality trend monitoring on mastered entities
- Supporting record
- Quality scorecard per data domain
How AI supports Master Data Management
Low-confidence matches are routed to stewards for decision.
Assistance drafts or summarizesRecommendation proposes for a person to acceptPredictive forecasts or flagsWorkflow-triggered runs only as configured
- Recommendation
AI matching
Probable matches are scored and proposed.
- Recommendation
Quality rule suggestions
Profiling results suggest candidate rules for stewards.
- Recommendation
Duplicate detection
Probable duplicates are proposed with a confidence score; uncertain ones go to stewards.
- Predictive
Quality scoring
Data quality is assessed across datasets to prioritize stewardship effort.
- Recommendation
Auto-classification
Classification and tags are suggested from record content for owner confirmation.
- Predictive
Data anomaly detection
Inconsistencies and outliers in mastered data are flagged for review.
- Recommendation
Relationship discovery
Implicit relationships between entities are proposed for hierarchy decisions.
- Predictive
Stale-data prediction
Records likely to become outdated are flagged for proactive review.
Why enterprises choose Nexus for master data management
Lineage per attribute
Every value in a golden record records the source and rule — or steward decision — it survived from, so trust is inspectable rather than assumed.
Stewards decide what rules cannot
Confident matches follow configured rules; uncertain ones become stewardship tasks with both records side by side, so automation never silently merges what a person should judge.
Mastered where it is used
Golden supplier, asset, location and organization records are consumed directly by Vendor Management, Asset Management / CMDB, IT Financial Management and Analytics Hub.
How Master Data Management connects
Part of the Data, Governance & Intelligence Function, with relationships across Nexus modules and the wider iShield portfolio.
Asset Management / CMDB
Shares trusted asset and location data.
Vendor Management
Uses golden supplier records.
IT Financial Management
Relies on consistent cost centres and entities.
Analytics Hub
Reports on consistent data.
Across the portfolio
What changes for the organization
One version of facts
Systems agree.
Fewer errors
Duplicates stop propagating.
Accountable data
Owners and stewards are named.
Reusable data
Consumers subscribe instead of copying.
Where this module is used
